Hello,
I think we have an route update/setup issue in our network that may need
some attention.
First of all, I don't see the root cause so maybe we should investigate
a little.
As you may have noticed, a technical bulldozer issue forced me to move
the yo2tm server to a VPS cloud machine, which is not a bad move by itself.
I put a gw with amprd on it and added a new new portal entry,
44.182.21.1 via 89.33.44.100. The route appeared in the RIP broadcasts a
few minute later and everything seems to be working as expected. Ping
was ok, connecting also.
So at the moment there are 2 overlapping subnets, 44.182.21.1/32 at the
new gw, and 44.182.21.0/24 at the old one.
Since I moved the call home map to the new location, too, I was
expecting that all nodes would show up in the new map, since all
notifications go to 44.182.21.1.
Normally, for routing purposes, the more precise route (/32) should take
precedence over the old one, and everything should be fine.
On the maps, there where some nodes appearing promptly via the new
gateway, like N1URO and some related radio nodes, but that was it. So
after 2 days, only 8 nodes switched to the new route, all others sending
their notification to the old address, WHICH SHOULD NOT HAPPEN.
The gateway was certainly updated as the internet pings hitting it moved
to the new target (some 20 public IPs).
Also, my netrom partners also followed after setting up the end axip
endpoints on the new machine.
I can not find an explanation for this, other than that there is some
route caching involved.
Maybe some of you have ideas/explanation on this behavior, and how to
fix it.
Marius, YO2LOJ